SOME RESULTS ON INCREMENTS
OF THE WIENER PROCESS
A. BAHRAM
Abstract. Let λ(T,aT ,α) =
n
h
2aT log
T
aT
+ α log log T + (1 − α) log log aT
io− 1
2
where 0 ≤ α ≤ 1 and {W (t), t ≥ 0} be
a standard Wiener process. This paper studies the almost sure limiting behaviour of
sup
0≤t≤T −aT
W (t)| as T −→ ∞ under varying conditions on aT and
1.
T
aT
λ(T,aT ,α) |W (t + aT ) −
.
Introduction
Let {W (t), t ≥ 0} be a standard Wiener process. Suppose that aT is a nondecreasing function of T such that
0 < aT ≤ T and aTT is nondecreasing. Csörgő and Révész [2], [3] etablished the following theorem.
Theorem 1.1. Let aT for T ≥ 0 satisfy
(1)
aT
(2)
0 < aT ≤ T,
aT
is nonincreasing.
T
(3)
is nondecreasing,
Received June 6, 2003.
2000 Mathematics Subject Classification. Primary 60J65.
Key words and phrases. Wiener process, increments of a Wiener process, a law of the iterated logarithm.
•First •Prev •Next •Last •Go Back •Full Screen •Close •Quit
Define βT = (2aT (log
(4)
T
aT
1
+ log log T ))− 2 . Then
lim sup
βT |W (T + aT ) − W (t)| = 1
sup
a.s.
T −→∞ 0≤t≤T −aT
(5)
lim sup
sup βT |W (t + s) − W (t)| = 1
sup
a.s.
T −→∞ 0≤t≤T −aT 0≤s≤aT
If, in addition,
(6)
lim
T −→∞
log
T
aT
log log T
= ∞,
then “limsup” may be replaced by “lim” in both equations (4) and (5).
Here and in the sequel we shall define for each u ≥ 0 the functions
Lu = log u = log(max(u, 1)),
and
L2 u = log log(max(u, e)).
ε stands for a positive number given arbitrarily, and C will be understood as a positive constant independent of
n, which can take different values on each appearance.
To simplify the notation, we will set
A(T, aT , α) =
sup
0≤t≤T −aT
B(T, aT , α) =
sup
λ(T,aT ,α) |W (t + aT ) − W (t)|,
sup λ(T,aT ,α) |W (t + s) − W (t)|,
0≤t≤T −aT 0≤s≤aT
•First •Prev •Next •Last •Go Back •Full Screen •Close •Quit
where
λ(T,aT ,α) =
− 12
T
2aT L
+ αL2 T + (1 − α)L2 aT
aT
2.
and
0 ≤ α ≤ 1.
Main result
In this section we shall investigate the analogous problem when βT is replaced by λ(T,aT ,α) . Our goal is to prove
the following result.
Theorem 2.1. Under assumptions (2) and (3) of Theorem 1.1, we have
(7)
lim sup A(T, aT , α) = 1
a.s.,
T −→∞
(8)
lim sup B(T, aT , α) = 1
a.s.
T −→∞
If we also have
(∗)
lim
T −→∞
L aTT
L((LT )α (LaT )1−α )
= ∞,
then
(9)
(10)
lim A(T, aT , α) = 1
a.s.,
lim B(T, aT , α) = 1
a.s.
T −→∞
T −→∞
Remark 2.1. Let us mention some particular cases .
1. For aT = T we obtain the law of iterated logarithm.
2. If α = 1, we obtain Csörgő-Révész theorem (see Theorem 1.1).
•First •Prev •Next •Last •Go Back •Full Screen •Close •Quit
3. If α = 0, under assumptions (2) and (3) of Theorem 1.1, then we also have
(11)
lim sup A(T, aT , 0) = 1,
a.s.,
T −→∞
(12)
lim sup B(T, aT , 0) = 1,
a.s.
T −→∞
If we also have lim
T −→∞
log
T
aT
log log aT
= ∞, then ” lim sup ” in Equation (11) and (12) may be replaced by “lim”.
Proof of Theorem 2.1. Our proof will be given in three steps expressed by the following three lemmas.
Lemma 2.1. Let aT be a nondecreasing function of T satisfying conditions (2) and (3) of Theorem 1.1. Then
for any ε > 0 we have
(13)
lim sup A(T, aT , α) ≥ 1 − ε.
T −→∞
Lemma 2.2. Let aT be a nondecreasing function of T satisfying conditions (2) and (3) of Theorem 1.1. Then
for any ε > 0 we have
(14)
lim sup B(T, aT , α) ≤ 1 + ε.
T −→∞
Lemma 2.3. Let aT be a nondecreasing function of T satisfying conditions (2), (3) of Theorem 1.1 and (∗)
of Theorem 2.1. Then for any ε > 0 we have
(15)
lim inf A(T, aT , α) ≥ 1 − ε.
T −→∞
Proof of Lemma 2.1. Let
C(T ) = λ(T,aT ,α) |W (T ) − W (T − aT )|.
•First •Prev •Next •Last •Go Back •Full Screen •Close •Quit
Using the well known probability inequality
2
2
1
1
x
1
x
1
√
exp −
(16)
− 3 exp −
≤ P (W (1) ≥ x) ≤ √
,
2
2
2π x x
2πx
for x ≥ 0, (see, e.g., [4, p.175]), it follows that
1−ε α 1−ε
aT
LaT
aT
P (C(T ) ≥ 1 − ε) ≥
≥
T (LT )α (LaT )1−α
T LaT
LT
1−ε LaT
aT 1−ε
aT
≥
≥
T LaT
LT
T LT
if T is big enough. We define the sequence {Tk } as follows: Let T1 = 1 and define Tk+1 by
Tk+1 − aTk+1 = Tk
if
ρ<1
and
Tk+1 = θk+1
if ρ = 1,
aT
where θ > 1 and lim
= ρ. The conditions (2) and (3) imply that aT is a continuous function of T and that
T →∞ T
ρ = 1 if and only if aT = T . Moreover T − aT is a strictly increasing function of T if ρ < 1. In the case ρ = 1 we
refer to the law of the iterated logarithm. So we assume that ρ < 1, (13) follows from
(17)
∞
X
k=2
aT
= ∞,
Tk (LTk )1−ε
as was shown in Csáki, Csörgő, Földes and Révész [1, Lemma 3.2], and the r.v. C(Tk ) (k = 1, 2, . . .) are
independent.
•First •Prev •Next •Last •Go Back •Full Screen •Close •Quit
Proof of Lemma 2.2. Let aTk = θk , θ > 1 and ε > 0. Using the inequality
P{
(18)
1
0
h− 2 |W (s) − W (s )| ≥ v} ≤
sup
0≤s0 ,s≤T,0≤s−s0 ≤h
CT
exp
h
−v 2
2+ε
,
where C is a positive constant depending only on ε (see in [2, Lemma 1∗ ]), we have
∞
X
P (B(Tk , aTk , α) ≥ (1 + ε))
k=1
∞
X
Tk
(1 + ε)2
Tk
exp{−2
(log
(LTk )α (LaTk )(1−α) )}
aTk
2+ε
aTk
k=1
ε 1+ε
∞ X
aTk
1
≤C
Tk
(LTk )α (LaTk )(1−α)
k=1
!1+ε
ε 1−α
∞ X
aTk
LTk
1
≤C
Tk
LaTk
LTk
k=1
1+ε
∞
ε
X
aTk
LTk
1
≤C
Tk
LaTk LTk
k=1
∞
ε
X aT
1
k
=C
<∞
Tk
(LaTk )1+ε
≤C
k=1
and an application of Borel-Cantelli Lemma gives
(19)
Notice that
lim sup B(Tk , aTk , α) ≤ 1 a.s.
k−→∞
•First •Prev •Next •Last •Go Back •Full Screen •Close •Quit
1≤
(20)
λ(Tk ,aTk ,α)
λ(Tk+1 ,aTk+1 ,α)
≤θ
if k is big enough. When Tk ≤ T ≤ Tk+1 , we have
lim sup B(T, aT , α) ≤ lim sup B(Tk+1 , aTk+1 , α)
T −→∞
k−→∞
λ(Tk ,aTk ,α)
λ(Tk+1 ,aTk+1 ,α)
≤ lim sup B(Tk+1 , aTk+1 , α) lim sup
k−→∞
k−→∞
λ(Tk ,aTk ,α)
λ(Tk+1 ,aTk+1 ,α)
.
Now choosing θ near enough to one, (14) follows from (19) and (20).
Proof of Lemma 2.3. We will set DT = {A(T, aT , α) ≤ 1 − ε}. Using inequality (18), for sufficiently large T ,
we have
P (DT ) ≤ P (
max
0≤i≤[ aT ]−1
λ(T,aT ,α) |W (i + 1)aT − W (iaT )| ≤ 1 − ε)
T
≤
1−
aT
α
T (LT ) (LaT )1−α
1−ε ![ aTT ]
ε
T
1
.
≤ 2 exp −
aT
(LT )α(1−ε) (LaT )(1−α)(1−ε)
Now, under condition (∗) and for all sufficiently large T ,
3−ε
T
≥ {(LT )α (LaT )1−α } ε .
aT
•First •Prev •Next •Last •Go Back •Full Screen •Close •Quit
Define Tk = eaTk = k.
Therefore
)
( 2α
∞
∞
∞
X
X
X
LT
k
P (DTk ) ≤ 2
exp{−(LTk )2α (LaTk )2(1−α) } = 2
exp −
(LaTk )2
LaTk
k=2
≤2
k=2
∞
X
k=2
exp{−(LaTk )2 } ≤ 2
k=2
∞
X
a−2
Tk = 2
k=2
∞
X
(Lk)−2 < ∞
k=2
which implies by Borel-Cantelli lemma that
lim inf A(Tk , aTk , α) ≥ 1 − ε, a.s.
(21)
k−→∞
When Tk ≤ T ≤ Tk+1 , we have aT − aTk ≥ 0 and by (3), it is easy to see that aT − aTk ≤
δ > 0. Thus
lim inf A(T, aT , α) ≥ lim inf
sup
λ(Tk+1 ,aTk+1 ,α) |W (t + aTk ) − W (t)|
T −→∞
aTk
Tk
≤ δaTk for any
k−→∞ 0≤t≤Tk −aT
k
− lim sup
sup
sup
T −→∞ 0≤t≤T −δaT 0≤s≤δaT
= lim inf
sup
k−→∞ 0≤t≤Tk −aT
k
− lim sup
sup
λ(T,aT ,α) |W (t + s) − W (t)|
λ(Tk ,aTk ,α) |W (t + aTk ) − W (t)|
sup
T −→∞ 0≤t≤T −δaT 0≤s≤δaT
λ(Tk+1 ,aTk+1 ,α)
λ(Tk ,aTk ,α)
λ(T,δaT ,α) |W (t + s) − W (t)|
λ(T,aT ,α)
.
λ(T,δaT ,α)
By Lemma 2.2 we have
(22)
lim sup
sup
sup
T −→∞ 0≤t≤T −δaT 0≤s≤δaT
λ(T,δaT ,α) |W (t + s) − W (t)| ≤ 1, a.s.
•First •Prev •Next •Last •Go Back •Full Screen •Close •Quit
We notice that
(23)
lim sup
T −→∞
λ(T,aT ,α)
= δ.
λ(T,δaT ,α)
The proof of Lemma 2.3 will be completed by combining (21), (22) and (23).
1. Csáki E., Csörgő M., Földes A. and Révész, P., How big are the increments of the local time of a Wiener process? Ann.
Probability 11 (1983), 593–608.
2. Csörgő M. and Révész P., How big are the increments of a Wiener process? Ann. Probability 7 (1979), 731-737.
3.
, Strong approximation in probability and statistics. Academic Press, New York (1981).
4. Feller W., An introduction to probability theory and its applications. Vol 2, 2nd . Willy, New York (1968).
A. Bahram, Laboratoire de Mathématiques, Université Djillali Liabès, BP 89, 22000 Sidi Bel Abbès, Algeria, e-mail: Abdelkader bahram@yahoo.fr
•First •Prev •Next •Last •Go Back •Full Screen •Close •Quit